This is the multi-page printable view of this section. Click here to print.
The Policy Library provides structured examples of Tetragon policies paired with relevant use cases. You can explore all documented policies, along with additional examples, in the policy library directory on Github.
For further community-maintained policy examples, refer to the policy examples directory on GitHub. These examples are valid policies provided by the community but they are not curated in terms of suitability for specific use cases or best practices.
Monitor execution of a binary in the /tmp directory.
Preventing execution of executables in /tmp is a common best-practice as several canned exploits rely on writing and then executing malicious binaries in the /tmp directory. A common best-practice to enforce this is to mount the /tmp filesystem with the noexec flag. This observability policy is used to monitor for violations of this best practice.
No policy needs to be loaded, standard process execution observability is sufficient.
jq 'select(.process_exec != null) | select(.process_exec.process.binary | contains("/tmp/")) | "\(.time) \(.process_exec.process.pod.namespace) \(.process_exec.process.pod.name) \(.process_exec.process.binary) \(.process_exec.process.arguments)"'
"2023-10-31T18:44:22.777962637Z default/xwing /tmp/nc ebpf.io 1234"
Monitor sudo invocations
sudo is used to run executables with particular privileges. Creating a audit log of sudo invocations is a common best-practice.
No policy needs to be loaded, standard process execution observability is sufficient.
jq 'select(.process_exec != null) | select(.process_exec.process.binary | contains("sudo")) | "\(.time) \(.process_exec.process.pod.namespace) \(.process_exec.process.binary) \(.process_exec.process.arguments)"'
"2023-10-31T19:03:35.273111185Z null /usr/bin/sudo -i"
Monitor execution of SUID “Set User ID” binaries.
The “Set User Identity” and “Set Group Identity” are permission flags. When set on a binary file, the binary will execute with the permissions of the owner or group associated with the executable file, rather than the user executing it. Usually it is used to run programs with elevated privileges to perform specific tasks.
Detecting the execution of setuid and setgid binaries is a common
best-practice as attackers may abuse such binaries, or even create them
during an exploit for subsequent execution.
Tetragon must run with the Process Credentials visibility enabled, please refer to Enable Process Credentials documentation.
No policy needs to be loaded, standard process execution observability is sufficient.
jq 'select(.process_exec != null) | select(.process_exec.process.binary_properties != null) | select(.process_exec.process.binary_properties.setuid != null or .process_exec.process.binary_properties.setgid != null) | "\(.time) \(.process_exec.process.pod.namespace) \(.process_exec.process.pod.name) \(.process_exec.process.binary) \(.process_exec.process.arguments) uid=\(.process_exec.process.process_credentials.uid) euid=\(.process_exec.process.process_credentials.euid) gid=\(.process_exec.process.process_credentials.gid) egid=\(.process_exec.process.process_credentials.egid) binary_properties=\(.process_exec.process.binary_properties)"'
"2024-02-05T20:20:50.828208246Z null null /usr/bin/sudo id uid=1000 euid=0 gid=1000 egid=1000 binary_properties={\"setuid\":0,\"privileges_changed\":[\"PRIVILEGES_RAISED_EXEC_FILE_SETUID\"]}"
"2024-02-05T20:20:57.008655978Z null null /usr/bin/wall hello uid=1000 euid=1000 gid=1000 egid=5 binary_properties={\"setgid\":5}"
"2024-02-05T20:21:00.116297664Z null null /usr/bin/su --help uid=1000 euid=0 gid=1000 egid=1000 binary_properties={\"setuid\":0,\"privileges_changed\":[\"PRIVILEGES_RAISED_EXEC_FILE_SETUID\"]}"
Monitor execution of binaries with file capabilities.
File capabilities allow the binary execution to acquire more privileges to perform specific tasks. They are stored in the extended attribute part of the binary on the file system. They can be set using the setcap tool.
For further reference, please check capabilities
man page; section File capabilities.
Detecting the execution of file capabilities binaries is a common
best-practice, since they cross privilege boundaries which make them a suitable
target for attackers. Such binaries are also used by attackers to hide their
privileges after a successful exploit.
Tetragon must run with the Process Credentials visibility enabled, please refer to Enable Process Credentials documentation.
No policy needs to be loaded, standard process execution observability is sufficient.
jq 'select(.process_exec != null) | select(.process_exec.process.binary_properties != null) | select(.process_exec.process.binary_properties.privileges_changed != null) | "\(.time) \(.process_exec.process.pod.namespace) \(.process_exec.process.pod.name) \(.process_exec.process.binary) \(.process_exec.process.arguments) uid=\(.process_exec.process.process_credentials.uid) euid=\(.process_exec.process.process_credentials.euid) gid=\(.process_exec.process.process_credentials.gid) egid=\(.process_exec.process.process_credentials.egid) caps=\(.process_exec.process.cap) binary_properties=\(.process_exec.process.binary_properties)"'
"2024-02-05T20:49:39.551528684Z null null /usr/bin/ping ebpf.io uid=1000 euid=1000 gid=1000 egid=1000 caps={\"permitted\":[\"CAP_NET_RAW\"],\"effective\":[\"CAP_NET_RAW\"]} binary_properties={\"privileges_changed\":[\"PRIVILEGES_RAISED_EXEC_FILE_CAP\"]}"
Monitor execution of the setuid() system calls family.
The setuid() and setgid() system calls family allow to change the effective user ID and group ID of the calling process.
Detecting setuid() and setgid() calls that set the user ID or group ID to root is a common best-practice to identify when privileges are raised.
The privileges-raise.yaml monitors the various interfaces of setuid() and setgid() to root.
jq 'select(.process_kprobe != null) | select(.process_kprobe.policy_name | test("privileges-raise")) | select(.process_kprobe.function_name | test("__sys_")) | "\(.time) \(.process_kprobe.process.pod.namespace) \(.process_kprobe.process.pod.name) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) \(.process_kprobe.function_name) \(.process_kprobe.args)"'
"2024-02-05T15:23:24.734543507Z null null /usr/local/sbin/runc --root /run/containerd/runc/k8s.io --log /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/log.json --log-format json --systemd-cgroup exec --process /tmp/runc-process2191655094 --detach --pid-file /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/2d56fcb136b07310685c9e188be7a49d32dc0e45a10d3fe14bc550e6ce2aa5cb.pid 024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392 __sys_setuid [{\"int_arg\":0}]"
"2024-02-05T15:23:24.734550826Z null null /usr/local/sbin/runc --root /run/containerd/runc/k8s.io --log /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/log.json --log-format json --systemd-cgroup exec --process /tmp/runc-process2191655094 --detach --pid-file /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/2d56fcb136b07310685c9e188be7a49d32dc0e45a10d3fe14bc550e6ce2aa5cb.pid 024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392 __sys_setgid [{\"int_arg\":0}]"
"2024-02-05T15:23:28.731719921Z null null /usr/bin/sudo id __sys_setresgid [{\"int_arg\":-1},{\"int_arg\":0},{\"int_arg\":-1}]"
"2024-02-05T15:23:28.731752014Z null null /usr/bin/sudo id __sys_setresuid [{\"int_arg\":-1},{\"int_arg\":0},{\"int_arg\":-1}]"
"2024-02-05T15:23:30.803946368Z null null /usr/bin/sudo id __sys_setgid [{\"int_arg\":0}]"
"2024-02-05T15:23:30.805118893Z null null /usr/bin/sudo id __sys_setresuid [{\"int_arg\":0},{\"int_arg\":0},{\"int_arg\":0}]"
Monitor creation of User namespaces by unprivileged.
User namespaces isolate security-related identifiers like user IDs, group IDs credentials and capabilities. A process can have a normal unprivileged user ID outside a user namespace while at the same time having a privileged user ID 0 (root) inside its own user namespace.
When an unprivileged process creates a new user namespace beside having a
privileged user ID, it will also receive the full set of capabilities. User
namespaces are feature to replace setuid and setgid binaries, and to allow
applications to create sandboxes. However, they expose lot of kernel
interfaces that are normally restricted to privileged (root). Such interfaces
may increase the attack surface and get abused by attackers in order to
perform privileges escalation exploits.
Unfortunatly, a report from Google shows up that 44% of the exploits required unprivileged user namespaces to perform chain privileges escalation. Therfore, detecting the creation of user namespaces by unprivileged is a common best-practice to identify such cases.
The privileges-raise.yaml monitors the creation of user namespaces by unprivileged.
jq 'select(.process_kprobe != null) | select(.process_kprobe.policy_name | test("privileges-raise")) | select(.process_kprobe.function_name | test("create_user_ns")) | "\(.time) \(.process_kprobe.process.pod.namespace) \(.process_kprobe.process.pod.name) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) \(.process_kprobe.function_name) \(.process_kprobe.process.process_credentials)"'
"2024-02-05T22:08:15.033035972Z null null /usr/bin/unshare -rUfp create_user_ns {\"uid\":1000,\"gid\":1000,\"euid\":1000,\"egid\":1000,\"suid\":1000,\"sgid\":1000,\"fsuid\":1000,\"fsgid\":1000}"
Monitor execution of the capset() system call.
The capset() system call allows to change the process capabilities.
Detecting capset() calls that set the effective, inheritable and permitted capabilities to non zero is a common best-practice to identify processes that could raise their privileges.
The privileges-raise.yaml
monitors capset() calls that do not drop capabilities.
jq 'select(.process_kprobe != null) | select(.process_kprobe.policy_name | test("privileges-raise")) | select(.process_kprobe.function_name | test("capset")) | "\(.time) \(.process_kprobe.process.pod.namespace) \(.process_kprobe.process.pod.name) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) \(.process_kprobe.function_name) \(.process_kprobe.args[3]) \(.process_kprobe.args[1])"'
"2024-02-05T21:12:03.579600653Z null null /usr/bin/sudo id security_capset {\"cap_permitted_arg\":\"000001ffffffffff\"} {\"cap_effective_arg\":\"000001ffffffffff\"}"
"2024-02-05T21:12:04.754115578Z null null /usr/local/sbin/runc --root /run/containerd/runc/k8s.io --log /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/log.json --log-format json --systemd-cgroup exec --process /tmp/runc-process2431693392 --detach --pid-file /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/9403a57a3061274de26cad41915bad5416d4d484c9e142b22193b74e19a252c5.pid 024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392 security_capset {\"cap_permitted_arg\":\"000001ffffffffff\"} {\"cap_effective_arg\":\"000001ffffffffff\"}"
"2024-02-05T21:12:12.836813445Z null null /usr/bin/runc --root /var/run/docker/runtime-runc/moby --log /run/containerd/io.containerd.runtime.v2.task/moby/c7bc6bf80f07bf6475e507f735866186650137bca2be796d6a39e22b747b97e9/log.json --log-format json --systemd-cgroup create --bundle /run/containerd/io.containerd.runtime.v2.task/moby/c7bc6bf80f07bf6475e507f735866186650137bca2be796d6a39e22b747b97e9 --pid-file /run/containerd/io.containerd.runtime.v2.task/moby/c7bc6bf80f07bf6475e507f735866186650137bca2be796d6a39e22b747b97e9/init.pid c7bc6bf80f07bf6475e507f735866186650137bca2be796d6a39e22b747b97e9 security_capset {\"cap_permitted_arg\":\"00000000a80425fb\"} {\"cap_effective_arg\":\"00000000a80425fb\"}"
"2024-02-05T21:12:14.774175889Z null null /usr/local/sbin/runc --root /run/containerd/runc/k8s.io --log /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/log.json --log-format json --systemd-cgroup exec --process /tmp/runc-process2888400204 --detach --pid-file /run/containerd/io.containerd.runtime.v2.task/k8s.io/024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392/d8b8598320fe3d874b901c70863f36233760b3e63650a2474f707cc51b4340f9.pid 024daa4cc70eb683355f6f67beda3012c65d64f479d958e421cd209738a75392 security_capset {\"cap_permitted_arg\":\"000001ffffffffff\"} {\"cap_effective_arg\":\"000001ffffffffff\"}"
Monitor the execution of binaries that exist exclusively as a computer memory-based artifact.
Often attackers execute fileless binaries that reside only in memory rather than on the file system to cover their traces. On Linux this is possible with the help of memfd_create() and shared memory anonymous files. Therefore, detecting execution of such binaries that live only in RAM or backed by volatile storage is a common best-practice.
Tetragon must run with the Process Credentials visibility enabled, please refer to Enable Process Credentials documentation.
No policy needs to be loaded, standard process execution observability is sufficient.
You can use the exec-memfd.py
python script as an example which will copy the binary /bin/true into
an anonymous memory then execute it. The binary will not be linked on the
file system.
jq 'select(.process_exec != null) | select(.process_exec.process.binary_properties != null) | select(.process_exec.process.binary_properties.file) | "\(.time) \(.process_exec.process.pod.namespace) \(.process_exec.process.pod.name) \(.process_exec.process.binary) \(.process_exec.process.arguments) uid=\(.process_exec.process.process_credentials.uid) euid=\(.process_exec.process.process_credentials.euid) binary_properties=\(.process_exec.process.binary_properties)"'
"2024-02-14T15:17:48.758997159Z null null /proc/self/fd/3 null uid=1000 euid=1000 binary_properties={\"file\":{\"inode\":{\"number\":\"45021\",\"links\":0}}}"
The output shows that the executed binary refers to a file descriptor
/proc/self/fd/3 that it is not linked on the file system.
The binary_properties
includes an inode
with zero links on the file system.
Monitor the execution of deleted binaries.
Malicious actors may open a binary, delete it from the file system to hide their traces then execute it. Detecting such executions is a good pratice.
Tetragon must run with the Process Credentials visibility enabled, please refer to Enable Process Credentials documentation.
No policy needs to be loaded, standard process execution observability is sufficient.
jq 'select(.process_exec != null) | select(.process_exec.process.binary_properties != null) | select(.process_exec.process.binary_properties.file != null) | "\(.time) \(.process_exec.process.pod.namespace) \(.process_exec.process.pod.name) \(.process_exec.process.binary) \(.process_exec.process.arguments) uid=\(.process_exec.process.process_credentials.uid) euid=\(.process_exec.process.process_credentials.euid) binary_properties=\(.process_exec.process.binary_properties)"'
"2024-02-14T16:07:54.265540484Z null null /proc/self/fd/14 null uid=1000 euid=1000 binary_properties={\"file\":{\"inode\":{\"number\":\"4991635\",\"links\":0}}}"
The output shows that the executed binary refers to a file descriptor
/proc/self/fd/14 that it is not linked on the file system.
The binary_properties
includes an inode
with zero links on the file system.
Audit BPF program loads and BPFFS interactions
Understanding BPF programs loaded in a cluster and interactions between applications and programs can identify bugs and malicious or unexpected BPF activity.
jq 'select(.process_kprobe != null) | select(.process_kprobe.function_name | test("bpf_check")) | "\(.time) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) programType:\(.process_kprobe.args[0].bpf_attr_arg.ProgType) programInsn:\(.process_kprobe.args[0].bpf_attr_arg.InsnCnt)"'
"2023-11-01T02:56:54.926403604Z /usr/bin/bpftool prog list programType:BPF_PROG_TYPE_SOCKET_FILTER programInsn:2"
Audit loading of kernel modules
Understanding exactly what kernel modules are running in the cluster is crucial to understand attack surface and any malicious actors loading unexpected modules.
jq 'select(.process_kprobe != null) | select(.process_kprobe.function_name | test("security_kernel_module_request")) | "\(.time) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) module:\(.process_kprobe.args[0].string_arg)"'
"2023-11-01T04:11:38.390880528Z /sbin/iptables -A OUTPUT -m cgroup --cgroup 1 -j LOG module:ipt_LOG"
Monitor loading of libraries
Understanding the exact versions of shared libraries that binaries load and use is crucial to understand use of vulnerable or deprecated library versions or attacks such as shared library hijacking.
jq 'select(.process_loader != null) | "\(.time) \(.process_loader.process.pod.namespace) \(.process_loader.process.binary) \(.process_loader.process.arguments) \(.process_loader.path)"'
"2023-10-31T19:42:33.065233159Z default/xwing /usr/bin/curl https://ebpf.io /usr/lib/x86_64-linux-gnu/libssl.so.3"
Monitor sessions to SSHd
It is best practice to audit remote connections into a shell server.
jq 'select(.process_kprobe != null) | select(.process_kprobe.function_name | test("tcp_close")) | "\(.time) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) \(.process_kprobe.args[0].sock_arg.family) \(.process_kprobe.args[0].sock_arg.type) \(.process_kprobe.args[0].sock_arg.protocol) \(.process_kprobe.args[0].sock_arg.saddr):\(.process_kprobe.args[0].sock_arg.sport)"'
"2023-11-01T04:51:20.109146920Z /usr/sbin/sshd default/xwing AF_INET SOCK_STREAM IPPROTO_TCP 127.0.0.1:22"
Monitor all cluster egress connections
Connections made outside a Kubernetes cluster can be audited to provide insights into any unexpected or malicious reverse shells.
PODCIDR=`kubectl get nodes -o jsonpath='{.items[*].spec.podCIDR}'`
SERVICECIDR=$(gcloud container clusters describe ${NAME} --zone ${ZONE} | awk '/servicesIpv4CidrBlock/ { print $2; }')SERVICECIDR=$(kubectl describe pod -n kube-system kube-apiserver-kind-control-plane | awk -F= '/--service-cluster-ip-range/ {print $2; }')jq 'select(.process_kprobe != null) | select(.process_kprobe.function_name | test("tcp_connect")) | "\(.time) \(.process_kprobe.process.binary) \(.process_kprobe.process.arguments) \(.process_kprobe.args[0].sock_arg.saddr):\(.process_kprobe.args[0].sock_arg.sport) -> \(.process_kprobe.args[0].sock_arg.daddr):\(.process_kprobe.args[0].sock_arg.dport)"'
"2023-11-01T05:25:14.837745007Z /usr/bin/curl http://ebpf.io 10.168.0.45:48272 -> 104.198.14.52:80"